All pull requests

Pull requests

• Ignore imaginary diagonal of C on input to herk/her2k

  #964·vera-rubin opened Oct 14, 2026·

  ·
  1

• sifive_{rvv,x280}: configure BLIS_SIMD_MAX_SIZE based on VLEN

  #963·myeh01 opened Oct 12, 2026Contributor·

  ·

• sifive_rvv: update packm kernels to work for any LMUL_{M,N}R

  #961·myeh01 opened Oct 10, 2026Contributor·

  ·

• dynamic block sizes with env. variables

  #960·stepannassyr opened Oct 7, 2026·

  ·

• fix off_m/off_n error in bli_gemm_ker_var2

  #959·stepannassyr opened Oct 7, 2026·

  ·

• rv64iv: stop compiling all reference kernels at -O0 with gcc

  #958·forhpc opened Oct 6, 2026·

  ·
  2

• sifive_rvv: fix packm LMUL_MR for VLEN < 512 (wrong results for column-major operands)

  #957·forhpc opened Oct 5, 2026·

  ·
  2

• rv64iv: 2vx8/8x8 gemm for VLEN≥256 (SpacemiT X100/A100)

  #955·hugomeiland opened Sep 27, 2026·3 tasks·

  ·
  3

• Add optional all-dims-small (AND) rule for SUP threshold

  #944·jeffhammond opened Jul 16, 2026Member·

  ·

• Multithread the memory-bound level-1/2 operations (OpenMP)

  #943·jeffhammond opened Jul 16, 2026Member·

  ·
  2

• Add optional single-thread cutoff for small level-3 problems

  #942·jeffhammond opened Jul 16, 2026Member·

  ·

• Prototype: weighted-static partitioning for heterogeneous cores

  #941·jeffhammond opened Jul 16, 2026Member·

  ·

• Add optional huge-page-backed memory pool

  #940·jeffhammond opened Jul 16, 2026Member·

  ·
  1

• Add cortexx925 and cortexa725 sub-configs for NVIDIA GB10

  #939·jeffhammond opened Jul 16, 2026Member·

  ·
  2

• Automatically detect hardware type for zen5 and later

  #927·edwsmyth opened Mar 30, 2026Contributor·

  ·
  4

• Amd zen4 support

  #925·harsdave opened Mar 26, 2026Contributor·

  ·
  1

• Implement optimized ccopyv and zcopyv kernels for Zen/2/3

  #924·harsdave opened Mar 24, 2026Contributor·

  ·
  2

• Add m4sme_p and m4sme_e sub-configurations

  #921·Luislo1 opened Mar 16, 2026·

  ·
  14

• gks: zero-initialize ref_cntx before use

  #883·jerryyiransun opened Aug 26, 2025·

  ·
  6

• Conditionally harden barriers.

  #872·devinamatthews opened Jun 22, 2025Member·

  ·

• sdgemm micro-kernel for the SG2042 RISC-V processor

  #868·evenieri opened May 6, 2025·

  ·
  2

• Proof-of-concept: speeding up gemm reference kernel

  #863·bartoldeman opened Mar 9, 2025Contributor·

  ·
  13

• Add skew-symmetric BLAS operations

  #805·devinamatthews opened Apr 26, 2024Member·

  ·
  21

• Implement "sandwich product" matrix operations

  #798·devinamatthews opened Feb 16, 2024Member·4 tasks·

  ·

• Armv8-A Row-major Kernel Improvements

  #698·xrq-phys opened Dec 16, 2022Collaborator·

  ·
  3